Abstract
AIM: To evaluate the sensitivity and specificity of transfesrrin dipstick test (Tf) in colorectal cancer (CRC) screening and precancerous lesions screening. METHODS: Eight hundreds and sixty-one individuals at high-risk for CRC were recruited. Six hundreds and eleven subsequently received the three fecal occult blood tests and colonoscopy with biopsy performed as needed. Fecal samples were obtained on the day before colonoscopy. Tf, immuno fecal occult blood test (IFOBT) and guaiac fecal occult blood test (g-FOBT) were performed simultaneously on the same stool. To minimize false-negative cases, all subjects with negative samples were asked to provide an additional stool specimen for a second test even a third test. If the results were all negative after testing three repeated samples, the subject was considered a true negative. The performance characteristics of Tf for detecting CRC and precancerous lesions were examined and compared to those of IFOBT and the combination of Tf, IFOBT and g-FOBT. RESULTS: A total of six hundreds and eleven subjects met the study criteria including 25 with CRC and 60 with precancerous lesions. Sensitivity for detecting CRC was 92% for Tf and 96% for IFOBT, specificities of Tf and IFOBT were both 72.0% (95% CI: 68.2%-75.5%; χ² = 0.4, P > 0.05); positive likelihood ratios of those were 3.3 (95% CI: 2.8-3.9) and 3.4 (95% CI: 2.9-4.0), respectively. In precancerous lesions, sensitivities for Tf and IFOBT were 50% and 58%, respectively (χ² = 0.8, P > 0.05); specificities of Tf and IFOBT were 71.5% (95% CI: 67.6%-75.1%) and 72.2% (95% CI: 68.4%-75.8%); positive likelihood ratios of those were 1.8 (95% CI: 1.3-2.3) and 2.1 (95% CI: 1.6-2.7), respectively; compared to IFOBT, g-FOBT+ Tf+ IFOBT had a significantly higher positive rate for precancerous lesions (83% vs 58%, respectively; χ² = 9.1, P < 0.05). In patients with CRC and precancerous lesions, the sensitivities of Tf and IFOBT were 62% and 69% (χ² = 0.9, P > 0.05); specificities of those were 74.5% (95% CI: 70.6%-78.1%) and 75.5% (95% CI: 71.6%-79.0%); positive likelihood ratios of those were 2.5 (95% CI: 2.0-3.1) and 2.8 (95% CI: 2.3-3.5). Compared to IFOBT alone, combining g-FOBT, IFOBT and Tf led to significantly increased sensitivity for detecting CRC and cancerous lesions (69% vs 88%, respectively; χ² = 9.0, P < 0.05). CONCLUSION: Tf dipstick test might be used as an additional tool for CRC and precancerous lesions screening in a high-risk cohort.